Skip to content

feat(services): run the remaining upstream preset units Dakota runs - #1285

Closed
hanthor wants to merge 1 commit into
chore/common-stale-overridesfrom
feat/dakota-upstream-preset-units
Closed

hanthor wants to merge 1 commit into
chore/common-stale-overridesfrom
feat/dakota-upstream-preset-units

Conversation

@hanthor

@hanthor hanthor commented Nov 16, 2026

Copy link
Copy Markdown
Member

What

Units are wired in 85-utah-desktop.preset, scripts/configure-services.sh and contracts/bluefin-desktop.toml [services]. ServiceMaskParityTests gets two new tests that check all three places agree, and the reference is docs/skills/desktop-contract/references/dakota-services.md.

Enabled, as on Dakota (measured: ghcr.io/projectbluefin/dakota:testing /etc/systemd/system/*.wants):

unit Dakota wants dir
nvmf-autoconnect.service, nvmefc-boot-connections.service default.target.wants
systemd-imdsd.socket sockets.target.wants
speech-dispatcherd.service multi-user.target.wants
podman.service default.target.wants
netavark-dhcp-proxy.socket / .service sockets.target.wants / default.target.wants
netavark-firewalld-reload.service firewalld.service.wants
netavark-nftables-reload.service nftables.service.wants

The two NVMe-oF units are each skipped by their own ConditionPathExists= unless /etc/nvme/{config.json,discovery.conf} exists (autoconnect) or /sys/class/fc/fc_udev_device/nvme_discovery exists (FC). netavark-firewalld-reload.service isn't named in #1272, but Dakota enables it and Utah runs firewalld, so it belongs with the rest of the netavark set.

Kept off, asserted in [services].disabled: mdcheck_start.timer, mdcheck_continue.timer, mdmonitor-oneshot.timer.

  • Utah keeps Fedora's raid-check.timer (parity(dakota): Utah-only system services Dakota doesn't run #1268 owner decision), which scrubs every array weekly. The mdcheck pair is a second scrub scheduler for the same arrays.
  • All three timers are WantedBy=mdmonitor.service. On Dakota that service is static and starts only when udev assembles an array (63-md-raid-arrays.rules). On Utah, Fedora's preset enables it at boot on every machine, so the timers would be pulled in everywhere. In fedora:44, mdadm --monitor --oneshot --scan with no mdadm.conf prints "No mail address or alert command - not monitoring." and exits 1, so the oneshot reminder would fail daily on non-RAID machines.

systemd-homed-firstboot.service (#1272) belongs to #1238, which carries the homed decision.

Stacked on #1283.

Closes #1271
Closes #1272

Verification

  • Every unit above exists in utah:testing (/tmp/dakota-scan/utah/system-units.tsv). The enabled ones are listed as disabled there, and the mdadm timers as well.
  • Read the nvme-cli, mdadm and mdmonitor-oneshot units in fedora:44 and ran the mdadm --monitor --oneshot --scan check above.
  • just check passes (1332 tests).

Enable what upstream presets turn on and a booted Dakota runs, in the
preset, configure-services.sh and the desktop contract: nvme-cli's NVMe-oF
boot connections (condition-gated on config or FC hardware),
systemd-imdsd.socket, the system speech-dispatcherd.service, podman.service,
and netavark's DHCP proxy and firewall-reload hooks.

mdadm's mdcheck_start, mdcheck_continue and mdmonitor-oneshot timers stay
off and are asserted disabled: Utah keeps raid-check.timer (#1268), which the
mdcheck pair would duplicate, and Fedora's always-on mdmonitor.service would
pull the oneshot reminder in daily on every machine, where it exits 1
without an mdadm.conf alert address. systemd-homed-firstboot belongs to
#1238.

Closes #1271
Closes #1272

Assisted-by: Claude Opus via GitHub Copilot
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@github-actions github-actions Bot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Nov 16, 2026
@github-actions

github-actions Bot commented Nov 16, 2026 •

Copy link
Copy Markdown
Contributor

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: danathar, hanthor

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@hivecommons-hive hivecommons-hive Bot added author/clanker A contributor's own hive authored this PR under their login no-dco A commit is missing Signed-off-by — run `git rebase --signoff` role/trusted Author is trusted for unattended merge (trusted-author role or registry/claim tier in require_tiers) stage/050-review-pending Review swarm requested; awaiting a verdict for the current head labels Nov 16, 2026
@hivecommons-hive hivecommons-hive Bot added stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act and removed stage/050-review-pending Review swarm requested; awaiting a verdict for the current head labels Nov 16, 2026
@hivecommons-hive

hivecommons-hive Bot commented Nov 16, 2026 •

Copy link
Copy Markdown
Contributor

Withdrawn: this PR is not back in your court after all. The failing checks (contract) are red on the base branch too; this change did not break them. There is nothing to fix here: the PR is back on the normal path, and the hive checks it again once the base branch is green.

@hivecommons-hive

Copy link
Copy Markdown
Contributor

Superseded by #1276 (earlier fix for #1271). The merge queue carries one PR per issue, so this PR is held out of it and returned to its author; close it if #1276 covers the change, or let the delegate decide.

@hivecommons-hive hivecommons-hive Bot added stage/050-review-pending Review swarm requested; awaiting a verdict for the current head and removed stage/072-returned-to-contributor Contributor PR: CI red, changes requested or conflicting; diagnosis posted, author must act labels Nov 16, 2026

@Danathar Danathar left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Stand-in review (the hive's reviewers are out of quota). Approving. It only adds enablements Dakota runs (the NVMe-oF units, skipped by their own conditions; systemd-imdsd.socket; speech-dispatcherd; podman/netavark), keeps the mdadm timers off with tests, and doesn't touch homed. systemd-homed-firstboot stays for #1238. systemd-imdsd.socket is a Varlink socket (Accept=yes) that starts systemd-imdsd@.service only per connection. The units that fetch metadata over the network are pulled in by the imds generator on cloud, not by this socket, so a desktop makes no requests. The red contract is the stack's. Before merge: none of these units is in the testsuite's "No failed systemd units at boot" VM allowlist, so a review-VM run should confirm systemd-imdsd.socket (SELinux under Fedora 44 policy) and speech-dispatcherd start cleanly.

@hanthor

hanthor commented Nov 16, 2026

Copy link
Copy Markdown
Member Author

Folded into the part-2 landing PR (fold/dakota-sq) with this change unchanged apart from rebase and conflict resolution. Closing here; review continues there.

@hanthor hanthor closed this Nov 16, 2026
@hivecommons-hive hivecommons-hive Bot removed role/trusted Author is trusted for unattended merge (trusted-author role or registry/claim tier in require_tiers) stage/050-review-pending Review swarm requested; awaiting a verdict for the current head labels Nov 16, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

approved Indicates a PR has been approved by an approver from all required OWNERS files. author/clanker A contributor's own hive authored this PR under their login no-dco A commit is missing Signed-off-by — run `git rebase --signoff`

Projects

None yet

Development

Successfully merging this pull request may close these issues.

parity(dakota): mdadm check timers and NVMe-oF units

2 participants